1

Size and Conflict in Town Meeting Democracy

An Honors Thesis for the University of Vermont Political Science Department

by Marc Laliberte

2

Table of Contents

Part 1: Introduction………………………………………………………………….…………………………..4

Part 2: Hypothesis and Previous Work…………………………………………..……………………15

Part 3: Methodology…………………………………………………………………………………...………29

Part 4a: Test 1 Findings…………………………………………………………...…………………………38

Part 4b: Test 2 Findings……………………………………...………………………………………………46

Part 5: Discussion………………………………………….…………………………………………………..52

Part 6: Works Cited………………………………………………………….…………………………………57

Appendix A……………………………………………………………………………….………………………..61

3

Tables and Figures

Table I: Descriptive Statistics for Test 1 Data……………….……………………………………..38

Table II: Correlations and Significance of all Test 1 Variables…………………………….…39

Table III: Multiple Regression Analysis of the Correlations of Discussion Time…..…45

Table IV: Descriptive Statistics for Test 2 Data…………………………………………………….46

Table V: Correlations and Significance of all Test 2 Variables……………………………….47

Figure I: Discussion Time and Size in Town Meetings………………………………………….40

Figure II: Discussion Time and Wealth in Town Meetings…………………………………….41

Figure III: Discussion Time and Budget Cost in Town Meetings……………………………41

Figure IV: Test 1 Correlations with Controls………………………………………………………..42

Figure V: Budget Costs and Majority Size in School Budget Votes………...……………....48

Figure VI: Town Size and Size of the Majority in School Budget Votes.………………….48

Figure VII: Town Size and Size of the Majority in School Budget Votes (2)…………….49

Figure VIII: Test 2 Correlations with Controls…………………………..………………………….50

Figure XI: Budgets Costs and the Size of Majority (No Unanimity Model)……………...61

Figure x: Town Size and Majority Size (No Unanimity Model)………………………………62

4

Part 1: Introduction

Students of government are often taught that the inspiration for America’s

democracy came from the government of classical Athens, where almost 3000 years

ago the people of this ancient civilization came together to govern themselves

according to the democratic principles of equality and justice.1 They learn that the

classical conception of government, with power resting in the people instead of an

arbitrary ruler, was the best historical example of legitimate sovereignty because of

the popular rule of the people. The surviving philosophical writings of Athenian

thinkers like Plato and Aristotle are also considered the basis of political theory that

inspired many other elements of our democracy. Clearly, modern democratic

governments owe a huge debt to the people of Athens and the other Greek city-

states for their wisdom and integrity in creating this form of government.

Students are also taught that the Founding Fathers who convened in

Philadelphia in 1787 created a Constitution that established and continues to guide

the American government to this day.2 They drew upon the ideas of preeminent

contemporary political philosophers of the Age of Enlightenment as well, such as the

social contract theory of John Locke3, and the system of checks and balances

conceived by Charles de Montesquieu4, to make an institution that is ruled by

majority while protecting individual liberty and the rights of minorities. The

American people are protected from tyranny of the majority and mob rule by an

1 Susan Welch et al. Understanding American Government: The Essentials (New York: Wadsworth Publishing, 2008), 53. 2 Welch et al. Understanding American Government, 58 3 John Locke, The Second Treatise of Government, (New York: Macmillan, 1986) 4 Charles de Montesquieu, The Spirit of Laws, trans. Thomas Nugent (London: G. Bell and Sons, 1878)

5

intricate balance of power between the three branches of government, and a

bicameral Congress, whereby all legislation is subjected to intense deliberation and

consideration of conflicting interests.5 This system has functioned reasonably well

for more than 2 centuries and Americans venerate the Constitution and the

Founding Fathers for their wisdom.

But how faithful has American government been to its democratic ancestry?

As we begin the 21st century in the extreme upper echelon of international power,

there is no doubt that the nation has experienced unparalleled economic and

military success. However, the political realm is unable to boast similarly

praiseworthy achievements. Of course, as a purportedly democratic institution,

judgment of American government must be based on the question of whether the

polity does what the people desire. The legitimacy of a democracy comes from the

power it gives its citizens, who ideally have control over the people making the laws

that govern them. A recent study by Public Policy Polling showed that Congress has

a disapproval rating of 85%, and that it is less popular than root canals and head

lice.6 The body often referred to as the People’s Branch of government has lost the

support of its namesake. Americans are clearly not pleased with the work their

national legislators are doing.

Perhaps the problem is that Congress is perceived to be doing nothing at all.

Congressional scholars noted in 2011, “The widespread consensus was that politics

5 Welch et al., Understanding American Government, 63 6 Tom Jensen, “Congress Less Popular than Cockroaches, Traffic Jams”, Public Policy Polling, Posted January 8th, 2013, http://www.publicpolicypolling.com/main/2013/01/congress-less-popular-than-cockroaches-traffic-jams.html

6

and governance were utterly dysfunctional.”7 Americans’ displeasure with Congress

and government is typically attributed to the rigid partisan polarization of the

institution.8 Democrats’ and Republicans’ inability to find common ground fills the

news on a daily basis. Although some have disputed the difficulty of passing

legislation in times of divided government9, the results of the 112th Congress show

that it was the least productive session in American history.10 Mann and Ornstein

blame much of the legislative conflict on the Republican Party’s confrontational

political tactics meant to halt the President’s policy agenda, although there are

certainly many other explanations.11 The two parties previously relied on

compromise facilitated by the more diverse political ideologies of their members,

but the regional realignment of the south after the Civil Rights Act caused each party

to become more ideologically cohesive and therefore less willing to cooperate.12 The

American democratic process needs legislators to work together to govern, and now

that they have lost the incentives for bargaining, uncompromising party politics

have brought the national government to a near standstill, where it is unable to

satisfy the demands of its citizens. This study is rooted in the belief that democratic

7 Thomas E. Mann and Norman J. Ornstein, It’s Even Worse Than it Looks: How the American Constitutional System Collided with the New Politics of Extremism, (New York: Basic Books, 2012), XII. 8 Mann and Ornstein, It’s Even Worse than it Looks, 44. 9 David R. Mayhew, Divided We Govern: Party Control, Lawmaking and Investigations, 1946-2002, (New Haven: Yale University Press, 2005). 10 Stephen Dinan, “Capitol Hill Least Productive Ever: 112th Fought ‘about everything,’” The Washington Post, January 9th, 2013, http://www.washingtontimes.com/news/2013/jan/9/capitol-hill-least-productive-congress-ever-112th-/?page=all. 11 Other explanations include the polarization of new forms of media like 24-hour cable news networks, gerrymandered legislative districts that make constituencies ideologically cohesive, primaries closed to nonpartisan voters that choose ideologically extreme candidates, and campaign finance reform opening the door to interest groups, among many more. Mann and Ornstein (103). 12 Mann and Ornstein, It’s Even Worse than it Looks, 47.

7

government is not working, and looks into the historical and theoretical literature

on the foundational elements of democracy for solutions to this problem.

At first glance it is clear how the political systems of ancient Greece inspired

the Founders to design a government that could be controlled by the people. After

shucking the authoritarian rule of England, under which they had been subjected to

laws and taxation that they had no say in making, it seems natural that the American

people would turn to a more democratic form of government through which they

could govern themselves. However, democracy in ancient Greece functioned

through intense deliberation in assemblies that all male citizens were invited to

attend.13 These assemblies considered and responded to public issues after debates

in which all members were allowed to speak, and judiciary bodies consisting of

randomly selected citizens settled public disputes.14 The Greek principle of isegoria

ensured that all individuals were welcome to speak their mind and make arguments

to show their peers the wisdom or dangers in different proposals. In addition,

scholars of ancient history argue that Cleisthenes, the Athenian leader typically

credited with initiating democratic rule, intended his reforms to resolve the intense

civic conflict among the powerful elite that was debilitating Athenian military

strength and public life.15 Unable to deal with political strife of the aristocracy,

Cleisthenes and the Athenians achieved stability by letting the people decide.

Democratic government began as a means of resolving conflict, not creating it.

Therefore the first obvious difference between American government and the Greek

13 R. Sealey, “Democratic Theory and Practice,” in The Age of Pericles, ed. Loren J. Samons II, (Boston: Cambridge University Press), 240. 14 Sealey, “Democratic Theory and Practice,” 244. 15 Sealey, “Democratic Theory and Practice,” 241.

8

conception of democracy is that the ancients relied on direct and intensive public

participation and deliberation from its citizens to resolve civic conflicts and make

the policies that governed the land.

The Founding Father’s incorporated a complex system of checks and

balances into the Constitution because they feared popular uprisings causing their

government to descend into mob rule. The legislative process is accordingly fraught

with places where a coalition or individual can delay or obstruct action. It has been

shown that “the American policy-making system of checks and balances and

separation of powers has more structural impediments to action than any other

major democracy.”16 Similarly, James Madison famously explained in Federalist No.

10 that a large republic of many varied interests grappling for power would prevent

any one faction from abusing the rights of others.17 This essay is considered

essential to understanding American government and is often said to represent the

genius of our democratic system. William F. Connelly and others claim that this

factionalism is what the Founders intended and describes the way the two-party

system functions in America today.18 The numerous and diverse interests of US

citizens have been condensed into two very equally balanced factions that are only

capable of preventing each other from acting. It appears that the American version

of democracy provides many opportunities for conflict to arise and depends on

groups’ inclination to opposition to thwart the tyranny of a majority.

16 Mann and Ornstein, It’s Even Worse than It Looks, 102 17 James Madison, “The Federalist No. 10,” in The Federalist: A Commentary on the Constitution of the United States, by Alexander Hamilton, James Madison, and John Jay, (New York: H. Holt and Company, 1898). 18 See William F. Connelly, James Madison Rules America: the Constitutional Origins of Congressional Partisanship, (Lanham: Rowman & Littlefield, 2010).

9

There is evidence that the Athenians were also aware of the vulnerability of

public assemblies to demagoguery and mob rule. Methods of persuasion and

rhetoric meant to sway the assembly were known as sophistry, which is the topic of

one of Plato’s well-known Socratic dialogues, the Gorgias. In this ancient text

Socrates talks with young students of sophistry who want to use it to gain power in

the assembly. He argues that it is sycophancy and therefore unjust, and places the

responsibility for diverting demagoguery on the individual.19 An Athenian trying to

take advantage of the public through persuasion of the assembly is acting

dishonorably, which does harm to his soul.20 Plato additionally claims that a citizen’s

personal attachment to his community should make him want what is best for it21,

thus distinguishing between giving the people what they want, “becoming their

servant and trying to please them,” and advocating difficult but beneficial laws

designed to make them better, “battling with the Athenians to make them as good as

possible.”22 The most honorable Greek statesmen, he says, earn their virtue by

working for the betterment of the community, not exercising their political power

for personal gain.

This theory has significant implications for this comparison of democratic

government in ancient Greece and America. First, the Athenians either found it was

unnecessary or just did not think to include procedural impediments to mob rule in

their original conception of democracy. The assembly functioned freely and

19 Plato, “Gorgias” in Gorgias, Menexenus, Protagoras, trans. Tom Griffith, (Boston: Cambridge University Press, 2009), 502d. 20 Plato, “Gorgias,” 522c-d. 21 Plato, “Gorgias,” 507e. 22 Plato, “Gorgias,” 521a.

10

supremely, deciding the best policies through deliberation and voting. Second, Plato

thought that honorable democratic decision-making was not easy and relied to

some extent on citizens’ emotional and moral interest in improving their

community. Clearly the Founders were more fearful of a tyrannical majority than

Plato and the ancient Greeks, and therefore included significant constraints on the

power of American government.

Finally, the most striking difference between these two forms of government

is their disparate sizes. The ancient Greek city-states were exactly what their name

suggests, small cities that controlled the public life within them and maybe some of

the surrounding hillsides. This demographic order is often attributed to Greece’s

natural geography, which made transportation over land difficult and therefore

restricted the power of organized government to its immediate surroundings.23

Fleck and Hanssen argue that democratic government emerged when the elite were

forced to allow agricultural landholders into the public decision-making process in

order to assure that the cities had access to enough resources.24 This suggests that

ancient Greek government was formed by the natural geography of Greece in

addition to the people who lived there. If democracy was developed in response to

the parameters of the landscape, then there is most likely a meaningful relationship

between the two. This means the modern form of democracy in America and many

other nations has been taken out of context.

23 Mogens Herman Hansen, Polis: An Introduction to the Ancient Greek City-State, (Oxford: Oxford University Press, 2006), 34. 24 Robert K. Fleck and F. Andrew Hanssen, “The Origins of Democracy: A Model with Applications to Greece,” Journal of Law and Economics, 49.1 (April 2006): 116.

11

Several scholars discuss this idea in their analyses of modern and historical

democracy. In Size and Democracy, Robert Dahl and Edward Tufte point out that the

political theory on government from the age of Aristotle and Plato up to the Age of

Enlightenment always considered democracy in its original form, the small city-

state.25 Prominent philosophers who greatly influenced the Founding Fathers, like

Rousseau and Montesquieu, saw more value in “the small, cohesive, highly

consensual city-state peopled by equal and substantially like-minded friends.”26

However, at around the time of the Constitutional Convention democratic theory

was adapted to fit the larger nation-state that was emerging as the dominant model

of statehood. The smaller city-state was too weak in confrontations with larger

nations; it did not have the resources and manpower to resist subjugation. The

allure of popular government and increased nationalism in the nation-states led to

the invention of political representation and then the conception of democracy as it

functions today.27 The American Constitution was a forerunner of this drastic

transformation in political thought, which made democracy possible in much larger

nation-states. This adaptation also stripped democracy of many of its worthwhile

elements, namely the direct participation of citizens in the political process and

their subsequent attachment to the community and interest in improving it. Thus it

appears as nations became larger and more able to conquer and annex the smaller

democracies, they then decided to take and bastardize their form of government.

25 Robert A. Dahl and Edward R. Tufte, Size and Democracy, (Stanford: Stanford University Press, 1973), 5-8. 26 Dahl and Tufte, Size and Democracy, 6. 27 Dahl and Tufte, Size and Democracy, 8-9.

12

Harvard scholar Jane Mansbridge calls this transformation of democracy the

“adversary revolution” that, in accord with Dahl and Tufte, “paved the way for

creating democratic institutions on a national scale.”28 Prior to this change in

political thought, democratic institutions like the Greek polis were based on

friendship and community and sought to determine the policies that would benefit

everyone the most. Members found equal pleasure in seeing their peers prosper as

themselves, so the goal of public decision-making was the common good.29

Mansbridge argues that the adversary revolution brought the acceptance and

institutionalization of conflict into the political process, in which democracy is now

a quantification and competition of many individuals’ self-interest.30 Evidence in

support of this theory is clearly seen in the current state of affairs in American

democracy.

Democracy was radically changed when it was expanded to the vast scale of

modern nations. It is widely recognized today that the representative form of

government used in many of these countries is not “real democracy,” in which

citizens are “called to meet in a deliberative, face-to-face assembly and to bind

themselves under laws they fashion themselves.”31 The real, original form of

democracy has been taken out of context and modified to fit the modern nation. As

conflict and disagreement bring American government to a halt, it is time to

reevaluate the effectiveness and functionality of this conception of democracy. The

28 Jane J. Mansbridge, Beyond Adversary Democracy, (New York: Basic Books, 1980), 5. 29 Mansbridge, Beyond Adversary Democracy, 13-14. 30 Mansbridge, Beyond Adversary Democracy, 16-17. 31 Frank Bryan, Real Democracy: The New England Town Meeting and How It Works, (Chicago: University of Chicago Press, 2004), 4.

13

philosophy that produced the system of checks and balances and Madison’s trust in

the factionalism of a large nation have effectively prevented tyrannical rule by

majority, but it has also undermined the ability of citizens to control their

government. This betrayal of the central tenet of democracy necessitates an

appraisal of the adaptations the Founding Fathers and other political philosophers

made to the government of the ancient Greeks.

Mansbridge’s theory of the institutionalization of conflict in democracy

suggests that these adaptations opened the door to conflict in democratic political

systems. Since these changes also facilitated the expansion of democracy to fit larger

states, this project will test the relationship between size and conflict in the political

systems of Vermont towns. These cases are useful because the town meeting form of

government provides the closest possible substitute for the conditions of

government in the ancient democracies of Greece. Bryan summarizes the

similarities and differences of these two forms of government clearly in Real

Democracy.32 Athenian democracy divided the region into demes that were close to

the size of Vermont towns, and each deme sent a delegation to the Council of Five

Hundred in Athens, akin to the state government in Montpelier.33 Although these

town governments are less autonomous than their Greek brethren because of the

federal bodies above them, their use of direct citizen participation and deliberation

make them appealing for this study. In addition, the relatively small size of these

communities is relevant and valuable to this study because of the argued

importance of smallness in fostering strong identity and cohesiveness in groups of

32 Bryan, Real Democracy, 4. 33 Bryan, Real Democracy ,10-11.

14

people.34 This study will explore the theoretical virtues of Greek democracy in a

quantitative analysis of size, conflict, and town meeting government, and investigate

the functionality of this form of government in larger communities to determine if

increased size has a negative effect on the character and effectiveness of democracy.

34 See Frank Bryan and John McClaughry, The Vermont Papers: Recreating Democracy on a Human Scale, (Post Mills, VT: Chelsea Green Publishing Company, 1989), 62-66, C.R. Hoffer, “Understanding the Community,” American Journal of Sociology, 36.4 (Jan. 1931): 623, and Aristotle, Nichomachean Ethics, trans. Roger Crisp, (Cambridge: Cambridge University Press, 2000).

15

Part 2: Hypothesis and Previous Work

The vices and virtues of small town New England are stereotypically based

on community intimacy; everyone knows each other’s personal and private lives,

New Englanders have intense pride in their town’s history, and a strong disdain for

outsiders (flatlanders).35 I believe that much of this familiarity and community-

mindedness comes from two factors. First, the percentage of people who live in

rural areas is much higher in northern New England states than anywhere else in

the country. Vermont (38.2%) and Maine36 (40.2%) had by far the lowest percents

of population living in urban areas in the 2000 U.S. Census.37 This means that a

larger proportion of New Englanders live in small towns than the rest of the

country. Second, the unique form of local government, town meeting, assembles

each town’s residents once a year to deliberate and democratically decide the

community’s affairs. I will demonstrate in this section that both of these factors

foster cohesiveness and a strong attachment to one’s municipality. I argue that these

qualities will influence the political decision-making process by causing citizens to

prioritize the town’s common interest, whereby they seek to improve the

community as a whole, over their individual self-interest. This study seeks to clarify

the role of size in democratic government and, more specifically, the ability of the

governments to resolve conflicting interests in their constituency. My hypothesis is

35 Bryan, Real Democracy, 124-127. 36 I personally consider these two states to be all that’s left of “real” New England. New Hampshire, the black sheep of the region, had an urban population of 59.3%, the 10th lowest in the country. In addition, the many other political, cultural, and demographic similarities between Vermont and Maine are very intriguing, but that is a subject for another day. 37 U.S. Census Bureau, “29 – Urban and Rural Population by State,” The National Data Book: The 2012 Statistical Abstract, last updated June 27, 2012,” http://www.census.gov/compendia/statab/cats/population.html.

16

that smaller towns will have less conflict in the local political process because the

citizens will be brought together by their mutual interest in making the community

better. The best example of this idea comes from Professor Frank Bryan, who told

me that whenever he wants to argue against another citizen at town meeting, he

always thinks of the next time he might need that person’s help pulling his tractor

out of the mud and usually reconsiders. Small towns breed forbearance in addition

to familiarity,38 because citizens know that they will encounter and most likely

require help from each other frequently throughout the year outside of town

meeting, and therefore tend to treat each other with respect. I believe the stronger

attachment residents of small towns have with their community makes the

democratic process more consensual and agreeable.

The theoretical support for this argument is rooted in the philosophies of the

original political theorists, Aristotle and Plato. Aristotle’s Politics studies the

different constitutional forms of the ancient Greece poleis in search of the ideal form

of government. His standard of value is the common good, as he states “those

constitutions which consider the common interest are right constitutions, judged by

the standard of absolute justice. Those constitutions which consider only the

personal interest of the rulers are all wrong.”39 With the remarkable prescience that

few but Aristotle can achieve, he then says that governments based on self-interest

“deviate from the true standard by not regarding the interest of all, and are thus

involved in a dilemma.”40 Whether Aristotle foresaw the idea of James Madison’s

38 Bryan, Real Democracy, 288. 39 Aristotle, Politics, trans. Ernest Baker (Oxford: Oxford University Press, 1946), 1279a. 40 Ibid.

17

factionalism that pits the citizens against each other, and the political conflict that

has consequently resulted41 is insignificant. He clearly believes the purpose of

government is finding the policies that are best for everyone, not weighing the

competing self-interests of individuals and enacting the heavier side. Common

interest is what brings people together to form societies and governments, Aristotle

claims, and it is through the pursuit of this noble good that everyone can have a

better life.42 In this piece he also makes the well-known assertion that states must

be small enough to allow all citizens to know “one another’s characters” and to

permit individuals to address the entire assembly by voice.43 Technology has

rendered this second boundary unnecessary for modern governments but the first

suggests that familiarity among the citizens of a polity is necessary for good

government, and the size of modern nation-states obviously precludes this

parameter. However, town meeting democracy, especially in the smaller towns of

Vermont, meets both of these conditions and additionally promotes the common

good through direct deliberation on matters concerning the community, which

Aristotle agrees is the better way to make decisions and prevent an aristocracy from

“enrich[ing] themselves from the public property”.44

Plato’s Gorgias shares this position on the precedence of the common good

over self-interest in government. He demonstrates that the ideal Athenian leader

should always pursue the policies that make the citizens better because this is the

41 "Partisanship and Stalemates Lead to Record Low Approval of Congress," CQ Almanac 2011, 67th ed., edited by Jan Austin, 1-3-1-8, Washington, DC: CQ-Roll Call Group, 2012, http://library.cqpress.com/cqalmanac/cqal-1390-77511-2461941. 42 Politics, Aristotle, 1278b. 43 Politics, Aristotle, 1326b. 44 Politics, Aristotle 1286a.

18

most just way to govern.45 His argument relies on a somewhat religious belief in a

day of reckoning after death46, but also defends the responsibility one has to his

community.47 Other scholars also discuss the importance of friendship in the Greek

conception of democracy. Horst Hutter describes Aristotle’s view of the relationship

between friends “as merely a more perfect expression of the friendship underlying

the political community as a whole.”48 Hutter contrasts this with the “individualizing

and alienating tendencies of modern society”49 to mark the contentious effects of

liberalism on the democratic process. In addition, he claims that democracy in

America was partly inspired by the fraternity of the New England Puritan

community, who designed the town meeting form of government being studied in

this paper.50 Friendship is important to the Greek conception of government

because people typically want to please their friends, i.e. do what is best for their

community. This prevented citizens from subjugating the common good to their

own private self-interest. Thus, where American government uses checks and

balances as well as factionalism to prevent mob rule, the Greeks relied on strong

community bonds to foster the pursuit of the public interest. This theory underlies

the hypothesis of this project. In smaller towns, citizens are much more likely to be

friendly with each other and therefore have one another’s interests in mind. This

uniformity of interests (everybody wants to do what is best for the community as a

whole) suggests there will be less disagreement in the political process.

45 Plato, Gorgias, 502c. 46 Plato, Gorgias, 526d. 47 Plato, Gorgias, 507e. 48 Horst Hutter, Politics as Friendship: The Origins of Classical Notions of Politics in the Theory and Practice of Friendship, (Waterloo, ON: Wilfrid Laurier University Press, 1978), 45. 49 Hutter, Politics as Friendship, 187. 50 Hutter, Politics as Friendship, 186.

19

Jane Mansbridge’s book Beyond Adversary Democracy shows many aspects

similar to this theory and investigates some ideas that are relevant to this study.

First, she makes a distinction between two concepts of democracy, unitary and

adversary. Unitary democracy is the form by which close-knit groups make

decisions consensually and for the benefit of everyone. This relies on the equality of

members and face-to-face deliberation, and she acknowledges that this is only

possible in small systems.51 The adversary form of democracy resulted from the

advent of liberalism and capitalism, and weighs the self-interests of citizens in order

to resolve conflicts.52 The adversarial revolution in democratic theory, which added

new concepts like representation to allow democracy to function on a larger scale,

“symbolizes, reinforces and institutionalizes division”53 and has caused citizens to

accept as fact that there will be opposition to their views. Mansbridge says that

modern democracy needs systems capable of performing both of these functions.

She attended consecutive town meetings in an unspecified Vermont town for 4

years to assess the ability of this form of democracy, and concludes that it is capable

of both the unitary and adversary modes of decision-making, but that “justice was

skewed and fairness corrupted by social coercion” taking place before the meeting.54

This is a fair criticism of town meeting democracy, but I believe that this pre-

meeting coalition forming is part of the deliberative process, and perhaps an

indication of the adversarial socialization of citizens. Beyond Adversary Democracy

supports my argument because it shows that conflict in the democratic process

51 Mansbridge, Beyond Adversary Democracy, 11-13. 52 Mansbridge, Beyond Adversary Democracy, 16-17. 53 Mansbridge, Beyond Adversary Democracy, 10. 54 Mansbridge, Beyond Adversary Democracy, 102-105.

20

ranges on a scale that depends on the personal relationships between citizens and

the strength of their identification with the community as a whole. Previous work on

this idea has made clear that “sense of community” is important in determining

political behavior, but these effects go in different directions depending on the social

context of the study.55 What Mansbridge ignores, and what I hope to demonstrate, is

that the causal variable in determining the attachment of individuals to their

community is size.

The problem of how to form a government that prioritizes the common good

over the self-interest of the most powerful elite or the most organized coalitions has

clearly been around for some time. Aristotle and Plato both suggest that putting the

most virtuous and enlightened man in the community in charge may solve this

dilemma, but that a man of this caliber is difficult to find.56 In the nearly 3000 years

since this question was first posed, political theorists have yet to come up with a

suitable answer. I argue that the solution is size. Empirical evidence to support this

thesis, that there is less conflict of self-interests in smaller communities, can be seen

in Size and Democracy, a study by Robert Dahl and Edward Tufte of how the size

variable affects several aspects of democratic government.57 First, one of the main

claims of this study is that in smaller democracies, individuals can see more clearly

that the common good is in their interest, while larger systems permit more

divergent views on what is good and on the proper goals of government.58 Diverse

55 Mary R. Anderson, “Beyond Membership: A Sense of Community and Political Behavior,” Political Behavior 31.4 (Dec. 2009): 603-627. 56 Aristotle, Politics, 1284b25, and Plato, The Republic, trans. R.E. Allen, (New Haven: Yale University Press, 2006). 57 Dahl and Tufte, Size and Democracy 58 Dahl and Tufte, Size and Democracy, 14.

21

views are acceptable and probably healthy for a deliberative body, but government

needs to be able to reconcile this conflict into coherent policy. Dahl finds evidence

showing that larger democracies tend to have a greater number of “organized

interests and interest groups” and their political processes are generally more

complex. 59 He also discusses the fact that in smaller communities dissenting actions

are more visible and there is more social pressure to conform to common norms,

but that this element of size would only be noticeable in communities smaller than

100,000 people.60 All of the towns being studied in this project are well below this

threshold, so this effect supports my hypothesis, that smaller towns will have less

conflict due to their shared common interest.

Dahl actually looks for evidence to support the relationship I am

investigating, and what he finds is kind of a mixed bag. First, he looks at the size of

legislative bodies in several democratic nations and American states and shows that

the quality of and participation in deliberation, and the capacity of citizen control

over government are diminished by larger size. He states that as size increases “the

parliament, in short, becomes less and less capable of functioning as an assembly.”61

Second, Dahl explains the results of a study on Swedish cantons that found the

voters in smaller cantons to have more homogenous preferences in elections.62 He

ultimately argues that as political power is centralized and ceded to global

institutions like the UN and the European Union to increase system capacity that

nations will have to improve local systems to compensate for diminished citizen

59 Dahl and Tufte, Size and Democracy, 39-40. 60 Dahl and Tufte, Size and Democracy, 42. 61 Dahl and Tufte, Size and Democracy, 80. 62 Dahl and Tufte, Size and Democracy 105.

22

effectiveness.63 In summary, Dahl & Tufte’s research suggests smallness in

democracy allows citizens more control over their government and tends to cause

more homogeneity of interests.

I believe that democratic government is not doing what it should, and that

what prevents it from doing this is that democracy was designed to govern vastly

smaller communities than it currently does. The academic community seems to

agree that this problem of scale creates a predicament for the theoretical and

practical foundations on which modern democracy is built, thereby causing its

ineffectiveness. Many of these scholars argue that the missing link in modern

democratic government is constructive public deliberation.64 In Strong Democracy

Benjamin Barber criticizes representative government and liberalism for creating

an individualistic political culture and reducing the participatory role of the citizen.

He notes that the concept of representation “rescues democracy from the problems

of scale” but has contradicted its democratic ideals by becoming a form of

oligarchy.65 The reforms he proposes “strengthen” democracy by providing better

opportunities for community deliberation, and he cites the New England town

meeting as a potential model for his proposals.66 Barber’s argument echoes a similar

tune to the theory on community cohesiveness that informs my hypothesis in this

study. He claims that the lack of civic engagement is the main problem of modern

63 Dahl and Tufte, Size and Democracy, 138. 64 For similar examples of this vein of scholarship, see Amy Gutmann and Dennis Thompson, Democracy and Disagreement, (Cambridge: Harvard University Press, 1996) and Robert Putnam, Bowling Alone: The Collapse and Revival of American Community, (New York: Simon & Schuster, 2000). 65 Benjamin R. Barber, Strong Democracy: Participatory Politics for a New Age, (Berkeley: University

of California Press, 1984), 290-291. 66 Barber, Strong Democracy, 272-273.

23

democracy. In the town meeting form of government, citizen deliberation plays a

central role, so my research will indicate how well democracy functions when direct

public discussion is involved. However, I believe that the qualities of deliberation

and democratic efficiency are both dependent on the size of the community.

Size and Democracy laments the lack of research on the correlates of size in

democratic nations, and implores political science to study this fundamental

relationship more.67 Frank Bryan noted almost 30 years later that this call was not

heeded.68 It is therefore very difficult to find academic research on the correlates of

size and aspects of democracy. Relevant comparative studies show that smaller

nations are more likely to form and sustain democratic governments, and this is

often attributed to the “spirit of fellowship and community” that smallness

supports.69 Lassen and Serritzlew found that an increase in size has a significant

negative effect on citizens’ internal political efficacy, which they define as “citizens’

belief that they are competent to understand and take part in politics.”70 These

findings both suggest that democracy works better in smaller systems, which I hope

to show in this project. One study of political polarization in survey responders’

views on public spending found that citizens of smaller democracies are more likely

to have divergent opinions than in larger nations.71 This finding contradicts my

67 Dahl and Tufte, Size and Democracy, 28. 68 Bryan, Real Democracy, 70. 69 Dag Anckar,“Small is democratic, but who is small?” Arts and Social Sciences Journal, 2012. 70 David Dreyer Lassen and Soren Serritzlew, “Jurisdiction Size and Local Democracy: Evidence on Internal Political Efficacy from Large-scale Municipal Reform” American Political Science Review 105.2 (May 2011): 238. 71 Erik Lindqvist and Robert Ostling, “Political Polarization and the Size of Government,” American Political Science Review 104 (Aug. 2010): 543-565.

24

hypothesis, so I will control for the burden of public spending in my trials to provide

for this possibility.

Academic research on town meeting government is similarly lacking. This is

very surprising because it is the only form of direct democracy that remains in the

world today.72 On town meeting day, the residents of a community are asked to

assemble and decide the questions of local government, which include matters such

as zoning laws, property taxes, school budgets, and the salary of the town

dogcatcher. This gives the people close control over their government as well as the

opportunity to learn and debate each other’s views. Any summary of the literature

on town meeting democracy would be remiss to begin with any other work than

Real Democracy by Frank Bryan. This immense project uses a database of 1,435

cases to investigate what causes political participation in Vermont towns. Bryan

identifies a strong negative correlation between the size of a town and the

percentage of its residents that participate in town meeting.73 He uses a corollary of

Anthony Downs’ classic voter’s paradox theory, which states that citizens choose to

participate in politics despite their knowledge of the extremely low chance that their

vote will make a difference in the outcome74, to argue that smaller towns have more

participation because the power of an individual’s vote increases with a decrease in

the size of the voting body. This investigation tests many other factors of town

meeting government to find the variables that effect attendance and verbal

participation, several of which are relevant to this study. There is a negative

72 Bryan, Real Democracy, 12-13. 73 Bryan, Real Democracy, 78-79. 74 Anthony Downs, An Economic Theory of Democracy, (New York: Harper and Row, 1957).

25

correlation between town size and the percentage of citizens who speak at town

meeting.75 Similarly, there is also a negative correlation between the size of the

meeting and the percentage of citizens who participate in discussion.76 So more

people participate in discussion in smaller towns and smaller meetings. More verbal

participation could be interpreted as more disagreement among the citizens, so

these findings suggest that my hypothesis, that smaller towns will have less conflict,

could be incorrect.

Other major studies of town meeting democracy include Jane Mansbridge’s

Beyond Adversary Democracy, discussed above, and The New England Town Meeting

by Joseph Zimmerman. In this book Zimmerman provides a comprehensive account

of the structural and historical aspects of town meeting in the 6 New England states.

His survey research concurs with Bryan’s finding that attendance in smaller towns

is higher77, and he also finds that the majority (78%) of town officers believe the

decisions made in town meeting are at least “good.”78 Zimmerman claims that

conflict most often emerges in town meeting when budget proposals threaten to

raise the town’s property tax rate.79 I will therefore control for the effects of budget

costs in this study. If there is a relationship between spending and conflict, its

influence will be removed from the correlation of size and conflict to ensure the

purity of my findings. Smaller investigations into the dynamics of town meeting

democracy show that socioeconomic status, a factor known to influence

75 Bryan, Real Democracy, 178. 76 Bryan, Real Democracy, 158. 77 Joseph Zimmerman, The New England Town Meeting: Democracy in Action, (Westport, CT: Praeger Publishing, 1999), 93. 78 Zimmerman, The New England Town Meeting 79 Zimmerman, The New England Town Meeting, 172.

26

participation in national elections,80 does not correlate with an individual’s

likelihood of attending town meeting.81 Town meeting participation is thus a forum

of socioeconomic diversity that allows citizens of all social classes equal status in

deliberation.

Many scholars have taken up investigations into the role of disagreement and

competition in democracy. Unfortunately all of this research appears to use conflict

as the independent variable and focus primarily on its effects on participation. This

study is more concerned with what causes disagreement in politics, but

nevertheless some of these findings are important. McClurg demonstrates that

exposure to divergent views has little effect on the individuals in the majority of the

given social context, but causes people in the minority to participate less frequently

in politics and political discussions.82 Given the current state of partisan

relationships in the U.S. this makes a seriously critical assessment of democratic

deliberation. With the factionalism of modern democracy in mind, disagreement

shuns the citizens with less support for their argument. It would follow that

government needs a way to bring people together in political participation and

refocus on finding agreement and common ground. I will determine if size is a factor

in this necessity. Other research shows that exposure to political competition cause

youth to become more interested and more likely to participate in politics later in

80 Richard Murray and Arnold Vedlitz, “Race, Socioeconomic Status, and Voting Participation in Large Southern Cities,” The Journal of Politics, 39 (1977): 1064-1072. 81 Victor DeSantis and David Hill, “Citizen Participation in Local Politics: Evidence from New England Town Meetings,” State & Local Government Review, 36.3 (2004): 166-173, and Vivian Scott Hixon, “The New Town Meeting Democracy: A Study of Matched Towns,” Ph.D. diss., Michigan State University, 1971. 82 Scott D. McClurg, “Political Disagreement in Context: The Conditional Effect of Neighborhood Context, Disagreement, and Political Talk on Electoral Participation,” Political Behavior 28.4 (Dec. 2006): (362).

27

life.83 This suggests, as other scholars have argued, that there may be some value in

disagreement in democratic government.

There is also a significant amount of sociological research that supports the

argument that smaller communities will have more solidary interests. Philip E.

Slater finds that larger groups are likely to be less stable, and smaller groups more

likely to inhibit the expression of disagreement and dissatisfaction.84 Edwin J.

Thomas indentified a negative relationship between group size and role consensus,

and group size and quality of work performance, meaning that smaller groups

tended to agree on the importance of their different roles more, and produce better

work than larger groups.85 Larger size has a negative effect on group cohesiveness,

which may be reflected in the bonds that make community and by conflict and

division in the political process.

In summary, ancient Greek political philosophy highlights the importance of

friendship and community in promoting the common good over individual’s self-

interest in government. By expanding the capacity of democracy to fit modern

nation-states, significant adaptations were made to the Greek conception of

government that removed the motives for citizens to consider the public interest

over their own selfishness. The supremacy of self-interest is seen in the adversary

nature of modern democracy, which has been theoretically established and

83 Julianna Sandell Pacheco “Political Socialization in Context: The Effect of Political Competition on Youth Voter Turnout,” Political Behavior 30.4 (Dec. 2008): 431. 84 Philip E. Slater, “Contrasting Correlates of Group Size,” Sociometry, 21.2 (Jun., 1958): 129-139. 85 Edwin J. Thomas, “Role Conceptions and Organizational Size,” American Sociological Review 24.1 (Feb. 1959): 30-37.

28

criticized for its alienating effect on the individual. Empirical research suggests

larger size causes disharmony in the political process, but the lack of significant

research on the effects of size on governance make this claim tentative. I will now

attempt to empirically demonstrate that size positively correlates with political

conflict by testing Vermont town meetings, a form of democracy very similar to the

ancient Greek model, to show that the extension of democratic government to large

states has effectively dissolved the components, such as popular rule and

community-mindedness, that made it such an attractive form of societal

organization.

29

Part 3: Methodology

There is considerable variation in the methods political scientists use to

measure conflict and disagreement in government systems, most likely because it is

an abstract idea that has numerous elements. It can be represented by argument,

division, gridlock, or even by events external to the political process like protests or

petitions. Studies measuring conflict appear to vary depending on the scale of

government under consideration. National scale investigations into partisan

polarization and gridlock in Congress look at the amount of significant legislation

passed,86 and the degree to which members vote with other members of their

party.87 Analyses of conflict in the American states use Presidential election results

to measure the extent of ideological disagreement.88 Since much of this information

is unavailable for smaller communities, studies at the local and neighborhood levels

rely on phone and mail surveys to determine conflict.89

Due to the lack of a widely accepted method of measuring political

disagreement, I have decided to test the relationship between size and conflict in

two separate ways. Both will use size as the independent variable, but

operationalize the dependent variable, the level of conflict in town meeting, in

different ways. Readily available data from Professor Frank Bryan provides a

detailed breakdown of the amount of time Vermont towns spent discussing various

proposals at town meeting. My first test compares the amount of argument on

86 Manabu Saeki, “Gridlock in the Government of The United States: Influence of Divided Government and Veto Players,” British Journal of Political Science 39.3 (July 2009): 593, and Mayhew, Divided We Govern. 87 Mann and Ornstein, It’s Even Worse than It Looks, 45, and Rosenthal and Poole, “The Polarization of American Politics.” 88 Pacheco, “Political Socialization in Context,” 422. 89 McClurg, “Political Disagreement in Context,” 355-356.

30

budget proposals to town size. Longer discussion will indicate more disagreement

and argument, and therefore more conflict in the town. Secondly, voting results

from town meeting decisions found in two Vermont newspapers’ reports on town

meeting day will be used to determine conflict. Results that are more equally

opposed (closer to 50-50) will clearly show the presence of more conflict. I decided

to perform two separate trials for a couple of reasons. First, the difficulty of

quantitatively defining and measuring an abstract idea like conflict could create

doubts about the integrity of this methodology, so two trials will give it additional

support to withstand criticisms. Second, I have already demonstrated the

importance of the democratic process in town meeting, and measuring just the

voting results (the outcomes) will ignore the aspects of the political transaction in

the assembly. Therefore, I measure the outcomes as well as the process to develop a

more comprehensive look into democracy in town meetings. I argue that both

operational definitions measure the dependent variable, conflict, in the political

process, because argument, reflected in discussion time, is equally indicative of

conflict as disharmony, which will be seen in the vote counts.

In Test 1 I correlate town size and the amount of time Vermont towns spend

discussing budget proposals at town meeting, and then control for the effects of

wealth and cost. Discussion time will serve as a measure of conflict because towns

that have more disagreement will spend more time arguing in town meeting. The

data on discussion time comes from a database provided by Professor Frank Bryan

at the University of Vermont. Bryan and his students collected this information to

measure the attendance, participation, and demographics of randomly selected

31

town meetings throughout the state from 1978 to 1995.90 Bryan coded this data to

show the length of and participation in discussion on individual warning items.91

This study will be restricted to cases held within 3 years of a national census in

order to ensure the accuracy of the population and wealth information. Historical

census information on Vermont towns comes from Vermont Indicators Online,92 an

invaluable database of demographic information provided by the University of

Vermont’s Center for Rural Studies. There are 94 cases that fit these criteria.

The dependent variable, conflict in the deliberative process, is measured as

the amount of time citizens discussed budget referendum proposed and voted on at

town meeting. Debate in Vermont town meetings can only be limited by a seconded

call to question and a two-thirds vote to end debate; any town resident can speak

for as long as he or she wants on any warning item, and when everyone is done the

item is put to question. Various types of budgets (highway, selectmen’s, general

fund, and school) from the database were used, unless the town allowed citizens to

vote by Australian ballot and therefore did not have a discussion. The Australian

ballot is a different voting system that requires towns to hold an informational

meeting before town meeting day to discuss warning items, and then open polls on

town meeting day for citizens to vote by paper ballot.93 It is generally used by larger

towns to allow citizens who are unable to attend a daylong meeting to participate in

90 See Bryan, Real Democracy. 91 Apparently breaking down 1,435 town meetings into individual voting items was a more time consuming task than he expected. He only made it through the alphabetically ordered database to the towns that begin with M. There is no possible statistical bias in the first half of the alphabet that I can think of. 92 “Vermont Indicators Online,” Compiled by The Center for Rural Studies at University of Vermont,

(accessed October 1st-December 5th, 2012). http://www.vcgi.org/indicators/. 93 Zimmerman, The New England Town Meeting, 84.

32

the political process, although it has been shown to reduce town meeting attendance

and participation,94 and attendance at the informational meetings is very slim.95

I define town size, the independent variable, as the number of voting-age

citizens in the town counted in the most recent census. I chose eligible voters

because this gives a more accurate representation of the size of the political

community could participate in the meeting, since adults and their incomes and

property are taken into consideration in the political process and children are not.

Wealth, the first control variable, is operationalized as the town’s GDP per

capita adjusted for the effect of inflation. Inflation data comes from the Bureau of

Labor Statistics Consumer Price Index.96 I chose wealth as a control variable

because of the well-established correlation between voter behavior and

socioeconomic status.97 For a similar reason my second control variable is cost,

operationalized as the amount of money in the proposed budget adjusted for

inflation and divided by the adult population of the town. Budget total numbers

came from Professor Bryan’s town meeting database. This gives an estimation of

how much each citizen expects to pay in property taxes if the proposal under

consideration is approved. I chose these two variables as controls because I believe

they will influence how much a citizen is willing to contribute to the community if

he/she disagrees with the budget proposal. The amount of money an individual

94 Bryan, Real Democracy, 97, 168. 95 In The New England Town Meeting, Zimmerman shows that attendance at informational town meetings ranges from 2.4 to 5.9 percent. 96 Bureau of Labor Statistics, “Consumer Price Index of Inflation,” ftp://ftp.bls.gov/pub/special.requests/cpi/cpiai.txt (last updated November 16th, 2012, accessed November 21st, 2012). 97 Richard Murray and Arnold Vedlitz, “Race, Socioeconomic Status, and Voting Participation in Large Southern Cities,” The Journal of Politics, 39 (1977): 1064-1072.

33

makes in addition to the amount he is being asked to spend could affect how he/she

participates in the deliberation on the budget proposal at town meeting. If so, I want

to control for these effects to clarify the relationship between size and conflict. In

summary, Test 1 will measure the relationship between size and the indicator of

conflict, discussion time.

Test 2 will look for a relationship between town size and conflict in vote

totals from Vermont town meetings, and then control for the effects of cost. Many

towns’ voting results are reported in several Vermont newspapers in the days

following town meeting day, held every year the first Tuesday in March, although

approximately 15 percent of towns have chosen to hold their meeting on the

Monday before, the Saturday after, or a different date for various reasons.98 I

gathered voting results from the archived microfilm records of two newspapers, the

Burlington Free Press and the Rutland Herald, available in the Bailey/Howe Library

at the University of Vermont. I chose these two daily newspapers because they

contained the best reports on the events of town meeting and together cover a

larger majority of towns in Vermont. The Free Press provided information on towns

in the north, northwest, and central regions of Vermont, and the Herald focused on

towns in the south. The data used for this test comes from these reports for the

years 1989 to 1992. I chose these years because they provided the only centralized

reporting on town meeting available that was closest to a census, in order to ensure

the accuracy of population figures. Town meeting reporting in these newspapers

98 Zimmerman, The New England Town Meeting, 85.

34

from the years surrounding 1980, 1970, and earlier decades typically do not include

a tally of votes taken at the meeting. Since 1990 town meeting attendance has been

in gradual decline,99 so I chose the years surrounding the 1990 census to maximize

data availability and participation.

I restricted the data in this test to the results of votes on school budgets

because they are consistently reported in the papers, in contrast to others that

towns are asked to consider, like highway and selectmen’s budgets. Results from

some towns were unavailable because they postpone school budget decisions to a

day later in the spring when state and federal contributions to the budget are more

complete.100 Budgets can be approved or rejected in Vermont town meetings in 4

different ways, by voice vote, hand/standing, paper ballot, or Australian ballot. After

deliberation is ended, the town moderator asks all those in favor to proclaim their

support by saying “aye” and then all opposed by saying “nay.” Unless the moderator

cannot determine the result by these vocal proclamations, or any voter calls for

division, the item is passed or turned down by this voice vote.101 In this test I

consider school budgets passed by voice vote to be unanimously approved by the

town, because rejecting the opportunity to quantify the vote signifies that voters in

the minority accept the assembly’s decision.

If there is adequate division in the town meeting the moderator can ask

citizens to show their vote and be counted either by raising hands or standing,

unless seven voters request that the vote be taken by paper ballot. If the results of a

99 Bryan, Real Democracy, 133-135. 100 Bryan, Real Democracy, 102. 101 Zimmerman, The New England Town Meeting, 90.

35

hand/standing vote or paper ballot on a school budget were reported in one of the

newspapers in the years 1989-1992, it is used as data in test 2. It should be noted

that the newspapers do not describe the outcomes of town meeting day in a uniform

way. In some cases, they provide detailed descriptions of the debate, budget

amount, voting method, and results, but in others they only say whether the budget

proposals passed and who was elected as town officials. There appears to be no

pattern behind this inconsistency; I believe it depends on what the reporter felt was

necessary to include. I can see no statistical bias that will come as a result and am

therefore confident that this data is a randomly gathered sample of town meeting

results. Nevertheless, I made certain to only include cases that specifically reported

the school budget amount, voting method, and decision of the town assembly.

The final way towns can make decisions in Vermont town meetings is by

Australian ballot. Towns do this in a number of different ways and use it to decide

different items like the election of town officers or budget proposals. Some open the

polls on ballot items during town meeting, and then decide other items

simultaneously in the traditional town meeting format, while others holds

informational meetings the night before and then vote by Australian ballot the next

day.102 The results of Australian ballots on school budget referendums are included

in test 2, in contrast with test 1. This is necessary because a large portion of towns

(approximately 28%)103 require budget referendum to be determined by Australian

ballot, and excluding them would decrease the size of the sample considerably. Also,

I argue that for the purposes of this test, Australian ballot results indicate the

102 Bryan, Real Democracy, 94. 103 Zimmerman, The New England Town Meeting, 87.

36

conflict in Vermont towns more or less equally to decisions made in the town

assembly because they include the opportunity for deliberation at the informational

meeting and legitimize the town’s decision. The data for the independent variable,

town size, comes from Frank Bryan’s database of historical Vermont demographic

information, and data for the control variable, cost, comes from the budget amount

specified in the newspaper reports.

I operationalize the dependent variable, conflict, as the size of the majority

that made the decision on the school budget vote. That means that test 2 will show

the relationship of size and conflict regardless of the decision, either approval or

rejection, of the town voters. For example if there were 75 “aye” votes on the school

budget, and 25 “nays”, I divided the number of “aye” votes by the total number of

votes, 100, and multiplied this value by 100 to turn it into a percent, to calculate the

size of the majority. This will show the extent to which citizens of the town agree

with each other on the school budget and approximate the level of conflict in the

town. Again, in this test I consider voice votes to be unanimous, meaning some cases

will have a majority of 100% in support or opposition. While in certain cases this

may not seem to be an accurate representation of the town’s vote, if the voters in the

minority cannot sufficiently argue their case during discussion to mobilize adequate

opposition and show division, or then neglect to call for division or fail to cause a

paper ballot, this indicates their resignation to the town’s decision as well as an

extremely low level of disagreement in the town.

The operational definition of town size, the independent variable, is the

number of registered voters in the town, because historical data on the amount of

37

eligible voters (which I use for the size variable in Test 1) was no longer available

online from the Vermont Center for Rural Studies when I compiled this data.

Professor Bryan was thankfully able to provide the amount of registered voters in

all Vermont towns in 1990, which provide a nearly identical substitute.104 I will also

test for the effects of cost, the control variable, on conflict in school budget votes

because this may influence citizens’ vote on the budget. If they feel the amount is too

high or too low for any reason, such as their children attend the local school or they

recently acquired heavily taxed property, this will affect their vote independently of

town size, so it must be controlled. Cost is operationalized as the budget amount

under consideration divided by the amount of eligible voters in the town, to

estimate approximately how much each citizen is expecting to have to contribute.

This is not an accurate estimate of the amount each citizen will be taxed because

school budgets are funded by property taxes that vary based on assessed value, and

many towns receive supplementary state and federal assistance, but it is unlikely

that a town meeting attendee would be able to calculate his or her exact share of the

budget before it is brought to a vote anyway, so this number is most likely similar to

what he or she estimates is her personal financial burden for the budget.

104 Bryan, Real Democracy, 65 note 16.

38

Part 4a: Test 1 Findings

Test 1 contained 94 cases. Descriptive information for this data can be seen

in table I and is summarized here. The dependent variable, discussion time on

budget items, had a minimum of 0 at a town meeting in Wells, Vermont in 1982, and

the maximum was 88 minutes from the 1987 meeting in Georgia. The mean

discussion time was 20.2 minutes. Size, the independent variable, had a minimum of

228 eligible voters in the town of Belvidere and a maximum of 3,762 voting-age

citizens in Bristol, and the mean size was 1,434. Inflation-adjusted town GDP per

capita varied from $11,693 in Brownington to $37,675 in Charlotte, and the average

was $20,032. The other control variable, adjusted budget cost per person, ranged

$36.13 in Bethel in 1982 to $3,374.87 in Sherburne (now known as Killington) in

1989, and the mean was $392.11.

Table I Descriptive Statistics for Test 1 Data

Variable N Minimum Maximum Mean

Time 94 0 88 20.22

Size 94 228 3,762 1,434

Town Wealth 94 $11,693 $37,675 $20,032

Budget Cost 94 $36.13 $3,374.87 $392.11

I tested the simple linear relationships between these four variables using

standard (product moment) coefficients. The correlation (Pierson’s r) and

significance of town size and discussion time, and the relationships between these

39

variables and the controls were all found in this manner and can be seen in Table II,

in which the significant relationships (p<.05) are underlined.

Table II Correlations and Significance of all Test 1 Variables

Size Time Wealth Budget Cost

Size

X .295 p= .004

.342 p= .001

-.150 p= .150

Time .295 p= .004

X .256 p= .013

.149 p= .152

Wealth .342 p= .001

.256 p= .013

X .246 p= .017

Budget Cost -.150 p= .150

.149 p= .152

.246 p= .017

X

This information shows the hypothesized positive relationship between size

and discussion time, correlated at a statistically significant level. Size and town

wealth are positively correlated, which means that citizens of larger towns tend to

have higher incomes. The positive relationship between wealth and time indicates

that wealthier towns also spend more time discussing budget items, which is

contrary to what I expected. For the most part, the budget cost control varied

insignificantly with the other variables, however there was a positive correlation

between the two control variables, which indicates the logical assumption that

wealthier towns tend to support higher budget costs per capita. Scatter plots (with

trend lines showing the hypothesized relationship between size and discussion

time) and the other relationships between the control variables and the dependent

variable can be seen in Figures I, II, and III. I have also included a causal guide to the

40

data, Figure IV, which maps the various correlations discussed above as well as the

relationship between the independent and dependent variables accounting for the

effects of the controls.

41

42

rBT=.149 (p=.152)

Figure IV

Test 1 Correlations with Controls

Wealth (W)

rSW=.342 (p=.001)

Size (S) Time

(T)

rSB=-.150 (p=.150)

Budget Cost (B)

rST=.295 (p=.004) rST.W=.228 (p=.028) rST.B=.324 (p=.002) rST.WB=.262 (p=.012)

rWB=.246 (p=.017)

rWT=.256 (p=.013)

43

To read the causal map, note that the letter in parentheses signifies the

variables in the relationships. So “r” and then the two letters signify the correlation

between two variables. The letters that come after the period indicate variables that

are under control. For example, rST is the simple correlation between size and time,

and rST.W represents this relationship after controlling for the effects of wealth. The

number in parentheses is the significance of the relationship. Figure IV shows the

significant positive correlations of size and time, size and wealth, wealth and time,

and wealth and budget cost. This means that the amount of time spent discussing

budget proposals at town meeting increases for larger towns and wealthier towns,

that larger Vermont towns tend to be more wealthy than smaller ones, and that

wealthier towns pay more per citizen on their budgets than smaller towns. There

was no significant relationship between size and budget cost or between budget

cost and discussion time. After controlling for wealth, the relationship between size

and discussion time is weakened but remains positive, and after controlling for

budget costs the correlation is strengthened. Controlling for both cost and wealth

slightly diminishes the strength of the relationship between size and discussion

time. This indicates that the relationship between size and discussion time

withstands the controls; it is still significant after removing the effects of wealth and

budget cost from data. Take for example the Charlotte highway budget discussion in

1988. Charlotte is a larger town, and the test indicates that it will therefore argue

longer about the budget. It is also a relatively wealthier town, which the model also

44

predicts will cause longer discussion time. Controlling for wealth takes away the

possibility that Charlotte’s long discussion time was caused by the town’s deep

pockets, and afterwards, since the relationship between size and discussion time is

still strong and significant, it is clear that Charlotte’s prolonged argument on the

budget was a result of its larger size. Based on this information, it can be concluded

that town size and wealth positively affect the amount of time spent discussing

budget proposals at town meetings, an indicator of political conflict in the town.

Figures I, II, and III show the lines of best fit for their respective correlations

as well as the percentage of the variance (R2) in the dependent variable (time)

explained by the independent and control variables (size, wealth, and cost). These

findings are all considerably lower than expected. Town size explains only about 9%

of the variance in discussion time spent on budget proposals, and wealth describes

about 7%. Again, there was no significant relationship found between budget costs

and time. I also administered a stepwise multiple regression on this data to gain

further understanding of these relationships. The results of this test can be seen in

Table III. The regression excluded the wealth control variable because it was

statistically insignificant in the equation. Nevertheless this analysis shows that the

combined effects of the size and budget cost variables significantly explain about

12.5% of the variance in town meeting budget discussion time. Accounting for the

cost of the budget to citizens increased the r-squared by about 4%.

45

Table III

Multiple Regression Analysis of the Correlates of Discussion Time (N=94)

Variable “r” R2 Increase in R2 Beta Sig. “p”

Size .295 .087 - .295 .004

Budget Cost .354 .125 .38 .324 .002

46

Part 4b: Test 2 Findings

Descriptive statistics on the data used for Test 2 can be found in Table IV

below. The minimum value for the size of majority variable was 50%, a school

budget vote in Thetford, Vermont in 1990 that managed to record an exact tie, 184-

184. The maximum size of majority recorded was 100%, from the 96 cases that

decided their school budget question with a unanimous voice vote, and the mean

majority was 77.9%. The amount of registered voters ranged from a minimum of 54

in Granby to the maximum of 5994 in Springfield. The mean town size was 1096

registered voters. The minimum budget cost per capita was $108.79 from

Greensboro in1992, and the maximum was $1,761.90 per town resident in the 1992

Bakersfield town meeting. The mean budget cost per capita was $745.05.

Table IV Descriptive Statistics for Test 2 Data

Variable N Minimum Maximum Mean

Majority Size 246 50% 100% 77.9%

Size 246 54 5994 1096

Budget Cost 246 $108.79 $1,761.90 $745.05

I used the same statistical analyses for Test 2 as Test 1 above. I calculated the

linear relationships between all three variables using standard (product moment)

coefficients. The correlation (Pierson’s r) and significance (p) of town size and size

of the majority, and the relationships between these variables and the control,

47

budget cost, were once again all found in this manner and can be seen in Table V, in

which the significant relationships (p<.05) are underlined.

Table V Correlations and Significance of Test 2 Variables

Size of Majority

Size Budget Cost

Size of Majority X -.564 p= .000

-.116 p= .070

Size

-.564 p= .000

X -.085 p= .186

Budget Cost

-.116 p= .070

-.085 p= .186

X

This information once more suggests that the hypothesized relationship

between size and conflict is correct. There is a strong negative correlation between

the town size and majority variables, which means that as town size increases the

group of citizens in majority agreement gets smaller. This indicates the presence of

heightened conflict in larger towns. The relationships between size and budget cost,

and budget cost and size of majority were both negative but not significant, which

means that the control variable did not have an effect on the relationship between

town size and size of majority as I expected. I have included Figure V to illustrate the

weak correlation between cost and majority size, and labeled some of the outlying

cases. Figure VI depicts the linear relationship between the amount of registered

voters and the size of the majority, and I have also included Figure VII to show the

logarithmic curve that represents this relationship. I believe the logarithmic curve

represents this trend better because it explains more of the variance in the data.

48

49

Furthermore, Figure VIII shows another causal map of the simple correlations of the

Test 2 variables including the effects of the control variable, cost. This is the same

format as the one used in Test 1, but with only one control.

Figure V shows the slightly negative correlation between budget cost and

size of the majority, as well as the low percentage of variance in majority explained

by the control. The R-squared value is only 1.3%, but remember the p value for this

correlation is .07 so this relationship is insignificant. This signifies the absence of a

relationship between these two variables. On the other hand, Figures VI and VIII

50

Figure VIII

Test 2 Correlations with Controls

Size (S) Majority (M)

show the strong relationship between the town size and size of majority variables.

The percentage of the variance explained by the linear relationship is 32%, and the

logarithmic equation describes 37%. The large cluster of cases at the top of each

scatter plot is the group of unanimous voice votes. This represents 98, or 39% of the

total 246 cases. Because of the likelihood that my methods will be criticized for

counting voice votes as unanimous decisions, I have included graphs of these

relationships excluding the cases with 100% size of majority in addendum.105

Without these voice votes, the relationship is very similar but the R-squared values

are slightly diminished. The causal map in Figure VIII displays the same

105 See Appendix A, page 61.

Cost (C)

rSC=-.085 (p=.186) rCM=-.116 (p=.070)

rSM=-.564 (p=.000) rSM.C=-.58 (p=.000)

51

relationships described above, in addition to the connection between town size and

majority size after controlling for the effects of budget costs. The correlation is

strengthened from -.564 to -.58. This means that the relationship survives the

control test intact. In summary, this information indicates that the size of a town has

a significant negative effect on the extent to which citizens agree with each other. In

larger systems of town government, the citizens tend to be more balanced in the size

of their opposing groups.

The final analysis of Test 2 data is a stepwise multiple regression, the results

of which are contained in Table VI. This shows the slight increase in percentage of

variance explained by accounting for the effects of budget cost. The R-squared value

increases from 32% to 35%.

Table VI Multiple Regression Analysis of the Correlates of Majority

(N=246)

Variable “r” R2 Change in R2 Beta Sig. “p”

Size -.564 .318 - -.564 .000

Budget Cost -.587 .345 .027 -.578 .000

52

Part 5: Discussion

The Vermont town meeting data suggests that the hypothesis was correct;

smaller communities have less conflict in local democratic government. This

relationship is seen in the amount of time they spend discussing proposals, a sign of

argument and hostility, and in the way they vote on matters of importance to the

community. The findings also withstand the controls done to test the possibility that

financial cost may affect conflict in addition to size. Cost of town budgets was not

correlated with conflict in either test, suggesting that citizens considered the

financial burden of their decisions a responsibility to their community more than a

personal affront. The context of community size therefore plays a large role in how

well real democracy functions to decide the common interest.

The opportunities for criticism I see in my study are the presumption that

voice votes indicate unanimity in Test 2, which I defend with the figures in Appendix

A showing that nearly the same relationship exists when they are removed from the

sample106, and the lack of more control variables. I hoped to submit these

correlations to more rigorous control tests, such as percentage population under 18

in the test on school budget votes, but the dearth of historical census data on

Vermont towns precluded these precisions. The database I was using, Vermont

Indicators Online, was taken down for lack of support from the University of

Vermont. My persistent entreaties for access to the information to the provider, the

Center for Rural Studies, were met with indifference.

106 In addition, Jane Mansbridge also considers voice votes to be a form unanimous, consensual decision-making in Beyond Adversary Democracy.

53

In Test 2, I prefer the curvilinear model for describing the relationship

between size and conflict for two reasons. First, the linear model places a definitive

boundary on the maximum size that allows democracy to make decisions at

approximately 4,000 registered voters. At this point it predicts the size of the

majority will diminish to 50%, meaning that perfectly balanced disagreement will

prevent the people from making a decision. The linear model also continues below

the 50% threshold, which does not fit into the methodological framework of the

study. Second, the effects of size on other aspects of democracy have previously

been shown to be concentrated in the extremely low end of the spectrum of town

size. Frank Bryan found that town meeting attendance decreased exponentially with

increases in size because of the diminishing returns of voter power as the size of the

electorate increases.107 Furthermore, Dahl notes “the characteristics we think

describe conflict in smaller systems are associated in a significant way only with

very small systems – with towns, say, having a population of under 10,000.”108 This

suggests that the important causative elements of size are only visible in the

smallest communities. In larger systems the relationship between size and conflict is

still valid; it is just unseen because the ability of community bonds to be meaningful

in large polities is already gone. The curvilinear model for Test 2 predicts the size of

the majority to shrink to a deadlock at around 6,000 registered voters, which

suggests that this is the upper limit of size for purposes of democratic government.

After this point conflict will prevent citizens from agreeing on the issues before

them, thus rendering popular rule and democratic government incapable.

107 Bryan, Real Democracy, 78-79. 108 Dahl and Tufte, Size and Democracy, 94.

54

This finding on the limit of democratic community size is clearly below the

population of all of the world’s nations today, but what about the communities in

which democratic government was born? Plato placed the size of the ideal polis at

5,040, citing the need for the city to have the resources to support itself, and to

allow all of the citizens to be familiar and friendly with each other.109 The actual

population of the Greek city-states is evidently the topic of considerable academic

debate.110 No one will ever know for sure, but archeological evidence allows an

informed guess. After surveying the many proposed methods of estimating the

population of ancient cities, Bowman and Wilson calculate that the average classical

Greek city-state had a population between 1,600 and 2,500 people.111 In his

exhaustive study of the available information on the ancient Greek polis, Hansen

concludes that over 80% of them had more than 1,000 citizens112 and exceptionally

large poleis (about 8% of the total) probably had up to 7,500 adult males113, the

Greek equivalent of the registered voters variable in this study. However,

democracy as we know it today is more closely associated with the way government

was practiced in the Athenian demes.114 These were smaller administrative units of

Athens that democratically decided local issues, where citizens could be heard by

their peers in a body that has been called “the center of their lives.”115 The

population of the demes has also been difficult for classics scholars to estimate, but

109 Laws, in The Dialogues of Plato, trans. By B. Jowett (New York, 1937), Vol. II: V, 738, 742; VI 771. 110 Bryan, Real Democracy, 8. 111 Bowman and Wilson, Settlement, Urbanization, and Population, 29. 112 This figure includes women and children, who could not participate in politics. Hansen, Polis (76) 113 Hansen, Polis, 83-84. 114 Bryan, Real Democracy, 8-9. 115 R. K. Sinclair, Democracy and Participation in Athens, (Cambridge: Cambridge University Press, 1988), 51.

55

most determine that the average size of the Athenian demes was less than 1000

people.116 The importance of community relationships was certainly heightened by

the small size of these units.

These assertions have a very interesting implication for this study because of

the similarity of Test 2’s prediction for the upper boundary of size in democracy and

the estimated size of the communities in which democracy was originally used. This

form of government seems to be inextricably reliant on the demographic aspects

that inspired it. It could be said that democracy sprung from the land – was a natural

occurrence resulting from the geographic, demographic, and cultural conditions of

the environment of the ancient Greek communities – and is therefore tied to this

specific niche in ancient history. Perhaps we are trying to make democracy work

outside of its natural habitat and have therefore condemned it from the start.

This research tells me that the ability of democracy to function properly

depends on the small size of the people it governs. In larger communities the strong

personal relationships between citizens that encourage them to act in the interest of

improving the polity as a whole break down, and instead the people fight to get the

best of one another. This theory argues that the factionalism of James Madison turns

democratic government on the scale of nation-states like the U.S. into a weak

shadow of its former standard, the virtues of which brought people together into

civilizations that showcased mankind’s capacity for empathy and cooperation rather

than reducing it to base creatures characterized by greed and hostility. Classical

116 See A.W. Gomme, The Population of Athens in the Fifth and Fourth Centuries B.C., (Oxford: Basil Blackwell, 1933) or John S. Traill, The Political Organization of Attica: A Study of the Demes, Trittyes and Phylai, and Their Representation in the Athenian Council, (Princeton, NJ: American School and Classical Studies at Athens, 1975).

56

Greek democracy worked because its people wanted the best for their community,

not for themselves. The determinant factor in bringing government closer to this

ideal is size.

Further research into the nature of democracy and size should assess this

theory on a larger scale, perhaps in the voting tendencies of the U.S. states, to see if

the correlation I found is visible at this level. I also wish there was a better way than

surveys to measure the extent to which voters act in their own interests or for the

common good. Developing a method to show this would clarify the effects of size on

the political behavior of individuals. Finally, it appears to me there will be an

incidental aura of apathy surrounding the findings of this project. Even if the

argument and empirical support are convincing, there will probably be a type of

“well, what can I do about it?” reaction because the modern democratic state is so

heavily institutionalized in the world. If nothing else, I hope that this project causes

people to revitalize the importance of their relationships in their public community

at the local level and beyond, because it is only through improved fellowship and

cohesiveness that we can make honorable progress as a civilization.

57

Part 6: Works Cited

Anckar, Dag. “Small is democratic, but who is small?” Arts and Social Sciences Journal 2012. Anderson, Mary R. “Beyond Membership: A Sense of Community and Political Behavior.” Political Behavior 31.4 (Dec. 2009): 603-627. Aristotle. Politics. trans. Ernest Baker. Oxford: Oxford University Press, 1946. Barber, Benjamin R. Strong Democracy: Participatory Politics for a New Age. Berkeley: University of California Press, 1984. Bryan, Frank. Real Democracy: The New England Town Meeting and How It Works. Chicago: University of Chicago Press, 2004. Bryan, Frank and John McClaughry. The Vermont Papers: Recreating Democracy on a Human Scale. Post Mills, VT: Chelsea Green Publishing Company, 1989. Bureau of Labor Statistics. “Consumer Price Index of Inflation,” last updated November 16th, 2012, accessed November 21st, 2012. ftp://ftp.bls.gov/pub/special.requests/cpi/cpiai.txt. Connelly, William F. James Madison Rules America: the Constitutional Origins of Congressional Partisanship. Lanham: Rowman & Littlefield, 2010. Dahl, Robert A. and Edward R. Tufte. Size and Democracy. Stanford: Stanford University Press, 1973. DeSantis, Victor and David Hill. “Citizen Participation in Local Politics: Evidence from New England Town Meetings.” State & Local Government Review. 36. No. 3 (2004): 166-173. Dinan, Stephen. “Capitol Hill Least Productive Ever: 112th Fought ‘about everything.” The Washington Post, January 9th, 2013. http://www.washingtontimes.com/news/2013/jan/9/capitol-hill-least- productive-congress-ever-112th-/?page=all. Anthony Downs. An Economic Theory of Democracy. (New York: Harper and Row, 1957). Fleck, Robert K. and F. Andrew Hanssen. “The Origins of Democracy: A Model with Applications to Greece.” Journal of Law and Economics 49.1 (April 2006): 115-146.

58

Gomme, A.W. The Population of Athens in the Fifth and Fourth Centuries B.C. Oxford: Basil Blackwell, 1933. Gutmann, Amy and Dennis Thompson. Democracy and Disagreement. Cambridge: Harvard University Press, 1996. Hansen, Mogens Herman. Polis: An Introduction to the Ancient Greek City-State. Oxford: Oxford University Press, 2006. Hixon, Vivian Scott. “The New Town Meeting Democracy: A Study of Matched Towns.” Ph.D. diss., Michigan State University, 1971. Hoffer, C.R. “Understanding the Community.” American Journal of Sociology 36.4 (Jan. 1931): 616-624. Horst Hutter. Politics as Friendship: The Origins of Classical Notions of Politics in the Theory and Practice of Friendsip. Waterloo, ON: Wilfrid Laurier University Press, 1978. Jensen, Tom. “Congress Less Popular than Cockroaches, Traffic Jams.” Public Policy Polling. Posted January 8th, 2013. http://www.publicpolicypolling.com/main/2013/01/congress-less-popular- than-cockroaches-traffic-jams.html. Lassen, David Dreyer, and Soren Serritzlew. “Jurisdiction Size and Local Democracy: Evidence on Internal Political Efficacy from Large-scale Municipal Reform.” American Political Science Review 105.2 (May 2011): 238. Lindqvist, Erik and Robert Ostling. “Political Polarization and the Size of Government.” American Political Science Review 104 (Aug. 2010): 543-565. Locke, John. The Second Treatise of Government. New York: Macmillian, 1986.

Madison, James. “The Federalist No. 10,” in The Federalist: A Commentary on the Constitution of the United States. By Alexander Hamilton, James Madison, and John Jay. New York: H. Holt and Company, 1898. Mann, Thomas E. and Norman J. Ornstein. It’s Even Worse Than it Looks: How the American Constitutional System Collided with the New Politics of Extremism. New York: Basic Books, 2012. Mansbridge, Jane J. Beyond Adversary Democracy. New York: Basic Books, 1980. Mayhew, David R. Divided We Govern: Party Control, Lawmaking and Investigations, 1946-2002. New Haven: Yale University Press, 2005.

59

McClurg, Scott D. “Political Disagreement in Context: The Conditional Effect of Neighborhood Context, Disagreement, and Political Talk on Electoral Participation.” Political Behavior 28.4 (Dec. 2006): (349-366). Montesquieu, Charles de. The Spirit of Laws. trans. Thomas Nugent. London: G. Bell and Sons, 1878. Murray, Richard and Arnold Vedlitz. “Race, Socioeconomic Status, and Voting Participation in Large Southern Cities.” The Journal of Politics, 39 (1977): 1064-1072. Pacheco, Julianna Sandell. “Political Socialization in Context: The Effect of Political Competition on Youth Voter Turnout.” Political Behavior 30.4 (Dec. 2008): 415-436. "Partisanship and Stalemates Lead to Record Low Approval of Congress." CQ Almanac 2011, 67th ed., edited by Jan Austin, 1-3-1-8. Washington, DC: CQ- Roll Call Group, 2012. http://library.cqpress.com/cqalmanac/cqal-1390- 77511-2461941. Plato. “Gorgias” in Gorgias, Menexenus, Protagoras. trans. Tom Griffith. Boston: Cambridge University Press, 2009. Plato. “Laws” in The Dialogues of Plato. trans. By B. Jowett. New York, 1937. Plato. The Republic. trans. R.E. Allen. New Haven: Yale University Press, 2006. Putnam, Robert. Bowling Alone: The Collapse and Revival of American Community. New York: Simon & Schuster, 2000. Saeki, Manabu. “Gridlock in the Government of The United States: Influence of Divided Government and Veto Players.” British Journal of Political Science 39.3 (July 2009): 587-607. Sealey, R. “Democratic Theory and Practice,” in The Age of Pericles, ed. Loren J. Samons II. Boston: Cambridge University Press. Sinclair R. K. Democracy and Participation in Athens. Cambridge: Cambridge University Press, 1988. Slater, Philip E. “Contrasting Correlates of Group Size.” Sociometry 21.2 (Jun., 1958): 129-139. Thomas, Edwin J. “Role Conceptions and Organizational Size.” American Sociological Review 24.1 (Feb. 1959): 30-37.

60

Traill, John S. The Political Organization of Attica: A Study of the Demes, Trittyes and Phylai, and Their Representation in the Athenian Council. Princeton, NJ: American School and Classical Studies at Athens, 1975. U.S. Census Bureau. “29 – Urban and Rural Population by State” in The National Data Book: The 2012 Statistical Abstract. last updated June 27, 2012, http://www.census.gov/compendia/statab/cats/population.html. “Vermont Indicators Online,” Compiled by The Center for Rural Studies at University of Vermont, (accessed October 1st-December 5th, 2012). http://www.vcgi.org/indicators/. Welch, Susan, John Gruhl, John Comer, and Susan M. Ridgon. Understanding American Government: The Essentials. New York: Wadsworth Publishing, 2008. Zimmerman, Joseph. The New England Town Meeting: Democracy in Action. Westport, CT: Praeger Publishing, 1999.

61

Appendix A

62